[1076] in SIPB_Linux_Development
ELF migration. (Go lemmings go!)
daemon@ATHENA.MIT.EDU (eichin@MIT.EDU)
Thu Oct 5 11:48:12 1995
From: eichin@MIT.EDU
Date: Thu, 5 Oct 1995 11:47:14 -0400
To: jered@MIT.EDU
Cc: linux-dev@MIT.EDU
In-Reply-To: <199510050307.XAA07601@vorlon.mit.edu> (jered@MIT.EDU)
I just want to overemphasize the comment that "we should not FORCE
users to switch to ELF." In spite of the fast number of ELF-only
programs out there, I'd recommend thinking in terms of a way to enable
users to run ELF software -- without "switching" to it per se.
If the spin on this is "here's a way to run some new programs, you
need to upgrade ld.so and add one directory" that's ok with me. I'd
actually discourage people from building ELF programs for quite a
while, though.
Perhaps I'm too sensitive to the fact that libc-5.2 has a broken
gethostbyname, and that the way libc-5.* has termcap integrated breaks
telnet pretty much completely.
I haven't looked at things closely enough to see how compatibility is
managed to quite understand how to avoid suffering (non minimize --
*avoid.* Totally.)
